Roles and Responsibilities:
Collaborative workbench: work closely with business units, data governance and data
engineer teams to get the business requirement and provide some ad-hoc analysis.
Working independently to identify the data issue, formulate the analysis framework, online
research and provide the result on time
Structured data cleansing and feature engineering: Develop data cleansing pipeline for high
volume data massaging and data exploration, including value fulfill rate checking, missing
value imputation, outliers’ detection and correlation checking
NLP: Explore and apply the state-of-the-art machine learning techniques to extract
information from unstructured data, like json, html files and web logfile.
Model development:
o Develop a variable selection pipeline to select out the features with
prediction/differentiation power.
o Model development: work closely with the teammate to develop machine learning
algorithms, like random forest, logistic regression, neural network, BERT and etc.
(supervised and unsupervised learning)
o Insights interpretation and have continuous integration/continuous delivery mindset
o End-to-end delivering the work, including the analysis documentation and
implementation code.
Some admin work, like requirement ticket raising to get data/business owner approval and
communicate/follow up with the process
Requirements:
Major in Data science, computer science, statistics, mathematics or equivalent experience — graduate degree preferable
Sound business knowledge in insurance industry will be a plus
Excellent analytical capability and understanding of various ML and analytical techniques
Proficiency in python or other coding language, like R, Scala and Spark
Familiarity with business visualization tools (e.g. PowerBI or Tableau)
Strong mathematics skills (e.g. statistic, algebra) and working knowledge of statistics
Experience with cloud solution such as Azure, Google and AWS will be a plus
Analytical with problem solving mindset, and ability to identify innovation opportunities,
define and deliver result